gait-capture-runpack
Capture and verify deterministic Gait runpacks from normalized run input. Use when asked to record a run, produce run_id or runpack artifacts, generate ticket-ready proof, or validate artifact integrity before handoff.
Best use case
gait-capture-runpack is best used when you need a repeatable AI agent workflow instead of a one-off prompt.
Capture and verify deterministic Gait runpacks from normalized run input. Use when asked to record a run, produce run_id or runpack artifacts, generate ticket-ready proof, or validate artifact integrity before handoff.
Teams using gait-capture-runpack should expect a more consistent output, faster repeated execution, less prompt rewriting.
When to use this skill
- You want a reusable workflow that can be run more than once with consistent structure.
When not to use this skill
- You only need a quick one-off answer and do not need a reusable workflow.
- You cannot install or maintain the underlying files, dependencies, or repository context.
Installation
Claude Code / Cursor / Codex
Manual Installation
- Download SKILL.md from GitHub
- Place it in
.claude/skills/gait-capture-runpack/SKILL.mdinside your project - Restart your AI agent — it will auto-discover the skill
How gait-capture-runpack Compares
| Feature / Agent | gait-capture-runpack | Standard Approach |
|---|---|---|
| Platform Support | Not specified | Limited / Varies |
| Context Awareness | High | Baseline |
| Installation Complexity | Unknown | N/A |
Frequently Asked Questions
What does this skill do?
Capture and verify deterministic Gait runpacks from normalized run input. Use when asked to record a run, produce run_id or runpack artifacts, generate ticket-ready proof, or validate artifact integrity before handoff.
Where can I find the source code?
You can find the source code on GitHub using the link provided at the top of the page.
SKILL.md Source
# Capture Runpack Execute this workflow to record an artifact safely and deterministically. ## Gait Context Gait is an offline-first runtime for AI agents that enforces tool-boundary policy, emits signed and verifiable evidence artifacts, and supports deterministic regressions. Use this skill when: - incident triage needs captured run artifacts for reproduction - CI gate failures need verifiable artifact identity and integrity - receipt/evidence generation depends on runpack metadata and digests Do not use this skill when: - Gait CLI is unavailable in the environment - no Gait run input or artifact source is available ## Workflow 1. Validate required input path: `<run_record.json>`. 2. Run record with JSON output (required): - `gait run record --input <run_record.json> --json` 3. Parse output fields: - `ok`, `run_id`, `bundle`, `manifest_digest`, `ticket_footer` 4. Verify artifact integrity (required): - `gait verify <run_id_or_bundle_path> --json` 5. Return a concise handoff block that includes: - `run_id` - `bundle` - `manifest_digest` - `ticket_footer` - verify status ## Safety Rules - Keep default capture mode as `reference`. - Do not switch to raw capture unless explicitly requested. - For replay workflows, prefer `gait run replay` (stub mode default); require explicit unsafe flags for real tool replay. - Do not invent `run_id`, digests, or verify results. - Treat non-zero exit from `gait run record` or `gait verify` as blocking errors. - Keep this skill wrapper-only: do not parse policy YAML or implement custom evaluator logic. ## Determinism Rules - Always use `--json` and parse structured fields. - Do not rely on text-only output for workflow decisions. - Keep output grounded in recorded artifact values only.